1. Alternative BPM Processing Scheme – May08 Demo
C.Perry – 08jan09

2. Signals from Striplines via LPF
Bunch
charge
~0.6e10
Peak V
400mV
Peak
spacing
(+ to -)
2.6ns
Bandwidth
~120MHz
(200MHz
+ 156MHz
filters in
series)

3. Signals from Striplines via LPF – Effect of Bandwidth
Bunch
charge
~0.6e10
Peak V
1250mV
Bandwidth
200MHz
Note how
rapidly
amplitude
varies with
bandwidth

4. Test Configuration phase- shifter Δ Hybrid Σ LPF LNA LPF to scope
156MHz
LNA
+20dB
LPF
200MHz to
scope
phase-
shifter
attenuator
-20dB
LPF
117MHz
splitter
from BPM
phase-
shifter Δ
Hybrid Σ
LPF
156MHz
LNA
+20dB
LPF
200MHz
splitter to
scope
phase-
shifter
attenuator
-20dB
LPF
117MHz
Hybrid: custom built composite unit
LNA: MiniCircuits ZX E: 20MHz-4GHz, +20dB, NF 3.9dB
Filters: MiniCircuits SBLP series
Phase shifters: Weinschel mechanical
Splitters: resistive
Main data runs: 3, of 50 pulses each, at 0.9E10 bunch charge (from BPM10, I think...)
Data statistically consistent, merged in the analysis

5. DAQ data - beam offset - unprocessed
top: sum, bottom: diff; X span 100ns; Y in mV uncorrected for gain/attenuation

6. DAQ data - beam offset – filtered and zeroed
top: sum, bottom: diff; X span 100ns; Y in mV uncorrected for gain/attenuation

7. DAQ data - beam offset – filtered and zeroed
left: sum, right: diff; X span 50ns; Y in mV uncorrected for gain/attenuation

8. DAQ data - beam centred - unprocessed
top: sum, bottom: diff; X span 100ns; Y in mV uncorrected for gain/attenuation

9. DAQ data - beam centred – filtered and zeroed
top: sum, bottom: diff; X span 100ns; Y in mV uncorrected for gain/attenuation

10. DAQ data - beam centred – filtered and zeroed
left: sum, right: diff;
X span 50ns; Y span 40mV (left), 25mV (right) (uncorrected for gain/attenuation )
Note quite small spread in bunch charge

11. DAQ data - beam centred – filtered and zeroed
left, right: processor difference outputs; centre: difference between the two
X span 20ns; Y span 25mV (uncorrected for gain/attenuation)
Position peak and range

12. Analysis
From data filtered and zero corrected (taking diff signal at fixed position):
sum: mean peak 19.8mV
diff signal: std dev 1.00mV each channel, 0.25mV difference between channels
assume equal noise from each channel => 0.18mV rms noise on diff output
Corrections:
20dB atten in sum => 198mV at output of hybrid
20dB gain in diff => 18uV noise referred to output of hybrid (measured gain: 20.2dB)
NB this is almost exactly the expected noise from the amplifier
hybrid design gives 3dB less sum output than an ideal device => ideal sum signal 280mV
hybrid gives 1.7dB loss at diff output => noise is equivalent to 22uV at ideal hybrid
NB the filters on sum and diff signals differed, but this is measured to be negligible
Equivalent position jitter:
expected position signal for 1mm offset from axis: (A-B)/(A+B) = 0.17
we have (noise in difference)/sum = 22uV/280mV = 7.8e-5
this is equivalent to a position jitter of 7.8e-5/0.17 = 4.6e-4mm = 0.46um
NB a beam jitter of about 3um rms is implied